Supporting tasks with adaptive groups in data parallel programming

O'Donnell, J.T. (2005) Supporting tasks with adaptive groups in data parallel programming. International Journal of Computational Science and Engineering, 1(2/3/4), pp. 86-98. (doi: 10.1504/IJCSE.2005.009694)

[img]
Preview
Text
odonnell3446.pdf

166kB

Publisher's URL: http://dx.doi.org/10.1504/IJCSE.2005.009694

Abstract

A set of communication operations is defined, which allows a form of task parallelism to be achieved in a data parallel architecture. The set of processors can be subdivided recursively into groups, and a communication operation inside a group never conflicts with communications taking place in other groups. The groups may be subdivided and recombined at any time, allowing the task structure to adapt to the needs of the data. The algorithms implementing the grouping and communications are defined using parallel scans and folds which can be executed efficiently in an abstract tree machine. This approach is best suited for massively parallel systems with fine grain processors.

Item Type:Articles
Keywords:Data parallel, task parallel, adaptive algorithms, processor groups, parallel scan, parallel tree machine.
Status:Published
Refereed:Yes
Glasgow Author(s) Enlighten ID:O'Donnell, Dr John
Authors: O'Donnell, J.T.
Subjects:Q Science > QA Mathematics > QA75 Electronic computers. Computer science
College/School:College of Science and Engineering > School of Computing Science
Journal Name:International Journal of Computational Science and Engineering
Publisher:Inderscience
ISSN:1742-7185
Copyright Holders:Copyright © 2005 Inderscience
First Published:First published in International Journal of Computational Science and Engineering 1(2/3/4):86-98
Publisher Policy:Reproduced in accordance with the copyright policy of the publisher.

University Staff: Request a correction | Enlighten Editors: Update this record